Brainstorm 2013 10 02

API Management

• API keys and rate limiting
• Versioning for API
• API client libraries in multiple languages

Ops & Deployment

• Write a publicly available deployment strategy guide
• Create AMIs, VMs, and Docker instances for people to spin up their own stack
• Provide automation help with Jenkins scripts, etc
• Work with hackers to identify vulnerabilities

Data & Datasets

• Access control layer for private data
  • Requires API keys or some sort of authentication
  • Would be nice to be as granular as particular fields in a slice
• Incremental data loads
• Scheduled, periodic data loads
• More advanced join/union SQL-like features
  • Could use similar caching strategy to aggregation
• Link concepts through multiple keys
• Remove datasets from the main repository
• Add "npm install"-like functionality for datasets
• Simplify dataset definition format
• Make a graphical dataset definition generator
  • Look at how chart.io loads data

Multiple Backends

• PostgreSQL backend
• Standard API for other backends

Admin Dashboard

• Track usage and hot queries
• Add new datasets through dashboard
• Some way for identifying useful indexes, maybe based on searching webserver logs for commonly used filters

Integration

• CKAN integration: CKAN for the front end and Qu for the back end
• Federation between Qu instances
• Federation between public and internal instances

End User/Presentation

• Set up a publicly available running instance of Qu with some interesting datasets
• Graphical query generator
• Make HTML templates customizable so that Qu instances can be themed
• D3-based data visualizations
• Show common queries/filters/aggregations to end users

Open Source

• Start using public Github for all issues, save those that are sensitive
• Break more components into external libraries
• Work on federal and open source evangelism